[Willrichardson]

For my university studies I’m needing to set up a StarCCM+ simulation where a cube achieves a drag coefficient of 1.05. I’ve tried many simulations with different models, mesh resolutions and wind tunnel sizes. I was wondering if anyone has done this and could inform me of their set up?
Thanks

[fluid23]

Whoever gave you that assignment gave you a very challenging problem. Star-CCM isn't the greatest tool for drag estimation. You can get valid results, but it can be challenging. However, a cube is a very problematic shape from a mesh quality standpoint.

First of all, what Re number do you need to achieve this? Do you know? How are you setting up the solver? Coupled/segregated? Steady/unsteady? Compressible? Lots of details are needed on what you tried and what went wrong for anyone to have a chance of helping you.

[cwl]

Originally Posted by fluid23

Star-CCM isn't the greatest tool for drag estimation.

I'm sorry for interfering - I might be missing some knowledge or specific experience, but .. why it isn't?

[fluid23]

Originally Posted by cwl

I'm sorry for interfering - I might be missing some knowledge or specific experience, but .. why it isn't?

No worries, it’s a question of personal experience. As far as the difficulties with star-ccm go, that’s years of accumulated experience in 3 or 4 different codes trying to do simple validations like circular cylinders, airfoils, etc. Star-ccm requires more work than other codes to get results that agree with published data, or at least that was true 5-10 yrs ago when I was doing that sort of thing more frequently.

I’m not saying it’s impossible, just more involved. I got better results on a cylinder and sphere validation on my first pass in SolidWorks Simulation than I did in Star-CCM.

That being said, star is the the most useful tool I’ve encountered in terms of available models, flexibility and capability.

As far as a square cylinder goes, that’s generally a more challenging problem than the classical cylinder or sphere problems. The sharp corners are a littler more challenging to mesh properly for a (presumably) inexperienced analyst, it’s an order of magnitude more difficult IMO. Circles are just easier to deal with using unstructured mesh than a square.

Again, not saying you CAN’T… just that it’s a more challenging task made more difficult for those just starting out on their path to mastering CFD.

[cwl]

Originally Posted by fluid23

Star-ccm requires more work than other codes to get results that agree with published data, or at least that was true 5-10 yrs ago when I was doing that sort of thing more frequently.

Can you please describe briefly - in which way?

Originally Posted by fluid23

I got better results on a cylinder and sphere validation on my first pass in SolidWorks Simulation than I did in Star-CCM.

Originally Posted by fluid23

As far as a square cylinder goes, that’s generally a more challenging problem than the classical cylinder or sphere problems. The sharp corners are a littler more challenging to mesh properly for a (presumably) inexperienced analyst, it’s an order of magnitude more difficult IMO.

That is completely understood - we all "love" angles and flow separation

Originally Posted by fluid23

Again, not saying you CAN’T… just that it’s a more challenging task made more difficult for those just starting out on their path to mastering CFD.

I'd say - in general drag is one of the hardest things for beginners despite it's seeming simplicity.

[fluid23]

The difficulties with star-ccm are a mix of personal preference and past experience. While a powerful tool, it isn't the most straightforward piece of software. The tree can be a little confusing, especially if you are expecting a top-down approach. The time it takes to set up a simple validation problem (like a cylinder) is (in my experience) anywhere from 20% to 50% longer than some simple software like Solidworks Simulation or Autodesk. Couple that with the fact that drag validation is challenging on a good day, and the odds start to stack up against you.

For an academic exercise, I think this person will spend as much or more time trying to figure out what they are doing wrong than they will actually learning something useful... and that may be true regardless of the software used.

And again, I will reiterate that I am not saying it's a bad piece of software, just more challenging than some others.